Purpose: This article uses an interactional analysis instrument to characterize patient-centered care in the primary care setting and to examine its relationship with health care utilization.Methods: Five hundred nine new adult patients were randomized to care by family physicians and general internists. An adaption of the Davis Observation Code was used to measure a patient-centered practice style. The main outcome measures were their use of medical services and related charges monitored over 1 year.Results: Controlling for patient sex, age, education, income, self-reported health status, and health risk behaviors (obesity, alcohol abuse, and smoking), a higher average amount of patient-centered care recorded in visits throughout the 1-year study period was related to a significantly decreased annual number of visits for specialty care (P ؍ .0209), less frequent hospitalizations (P ؍ .0033), and fewer laboratory and diagnostic tests (P ؍ .0027). Total medical charges for the 1-year study were also significantly reduced (P ؍ .0002), as were charges for specialty care clinic visits (P ؍ .0005), for all patients who had a greater average amount of patient-centered visits during that same time period. For female patients, the regression equation predicted 15.47% of the variation in total annual medical charges compared with male patients, for whom 31.18% of the variation was explained by the average percent of patient-centered care, controlling for sociodemographic variables, health status, and health risk behaviors.Conclusions